Impact of Negative Pressure Wound Therapy on Outcomes Following Pancreaticoduodenectomy: An NSQIP Analysis of 14,044 Patients

Background: Despite ongoing efforts to improve the pancreaticoduodenectomy technique and perioperative care, surgical site infection (SSI) remains a contributor to morbidity. Efforts to reduce SSI include the use of negative pressure wound therapy (NPWT), but studies and meta-analyses have been met with conflicting results. We aimed to provide an up-to-date large-scale cohort study to assess the impact of NPWT on SSIs. Methods: Utilizing the National Surgical Quality Improvement Program database, we included patients undergoing a pancreaticoduodenectomy between 2017 and 2021 and divided patients into the NPWT and non-NPWT cohorts. A bivariate analysis was performed to compare baseline characteristics and complication rates between the cohorts. Multivariate logistic regression analysis was performed to assess the independent effect of NPWT on 30-day serious complication, 30-day mortality, and the development of deep or superficial SSI. A priori sensitivity analyses were performed in high-risk and malignancy cohorts. Results: Of the 14,044 included patients, 1689 (12.0%) patients had a prophylactic NPWT device, while 12,355 (88.0%) did not. Patients were more likely to have NPWT if they had higher ASA scores, had diabetes, were dialysis-dependent, or had a hard pancreas, but they were less likely if they were a smoker, had steroid use, or had a bleeding disorder. Most complications occurred similarly between the two cohorts, including superficial and deep SSI, but NPWT patients had a longer length of stay (10.4 d vs. 9.5 d, p < 0.001) and higher organ space SSI (22.6% vs. 17.4%, p < 0.001). Following multivariable modeling to control for demographic differences, NPWT was not independently associated with a difference in likelihood of SSI (aOR 0.94, p = 0.691) or serious complications (aOR 0.958, p = 0.669). Furthermore, the sensitivity analyses of both high-risk and malignant subgroup also did not see an independent association of NPWT on the rate of SSI (aOR 0.98, p = 0.898 and 0.96, p = 0.788, respectively). Conclusion: NPWT is used infrequently and is not significantly associated with improved outcomes including in the high-risk or malignant subgroups based on multivariable analysis for surgical site infections nor did it improve the outcomes of 30-day serious complications in these subgroups. Considering this and other studies showing the limited benefit of NPWT in all-comers and in high-risk cohorts, it remains unclear whether NPWT offers benefits following PD.